High-income does not protect against hurricane losses

Damage due to tropical cyclones accounts formore than 50%of allmeteorologically-induced economic losses worldwide. Their nominal impact is projected to increase substantially as the exposed population grows, per capita income increases, and anthropogenic climate changemanifests. So far, historical losses due to tropical cyclones have been found to increase less than linearly with a nation’s affected gross domestic product (GDP). Herewe show that for theUnited States this scaling is caused by a sub-linear increase with affected populationwhile relative losses scale super-linearly with per capita income. Thefinding is robust across amultitude of empirically derived damagemodels that link the storm’s wind speed, exposed population, and per capita GDP to reported losses. The separation of both socio-economic predictors strongly affects the projection of potential future hurricane losses. Separating the effects of growth in population and per-capita income, per hurricane losseswith respect to national GDP are projected to triple by the end of the century under unmitigated climate change, while they are estimated to decrease slightly without the separation.